Table 3.

Results for two-class classification, part II.

Technique	Specificity	Specificity Avg | Std	Sensitivity	Sensitivity Avg | Std
FSL proximity-based	89.6–91.0%	90.4% | 0.5%	88.9–90.7%	89.9% | 0.6%
Softmax-based classification	89.0–90.2%	89.4% | 0.5%	88.9–90.7%	89.9% | 0.6%
FSL + XGBoost	89.1–90.7%	89.8% | 0.6%	86.5–88.5%	87.7% | 0.8%
FSL + Random Forest	89.4–91.9%	90.3% | 1.0%	86.3–90.1%	88.2% | 1.3%
FSL + Decision Tree	86.4–89.7%	87.6% | 1.2%	82.8–87.5%	85.0% | 1.8%
FSL + KNN − 5 neighbors	89.6–92.7%	90.8% | 1.1%	87.1–91.2%	88.8% | 1.4%
FSL + KNN − 20 neighbors	89.4–93.9%	91.6% | 1.7%	86.6–92.6%	89.8% | 2.2%
FSL + SVM with linear kernel	89.5–93.5%	91.6% | 1.5%	87.4–92.9%	90.3% | 1.8%
FSL + SVM with polynomial kernel	89.2–93.7%	91.0% | 1.6%	85.5–92.3%	88.2% | 2.3%
FSL + SVM with RBF kernel	90.0–93.5%	91.7% | 1.3%	88.1–92.8%	90.5% | 1.6%
FSL + SVM with Sigmoid kernel	68.2–93.4%	87.2% | 9.6%	66.2–92.2%	85.3% | 9.6%